I've got a project I want to work on, which involves an LCD, MAX232, and MSP430. The LCD requires 5V for logic and 3V for the backlight. The MAX232 will work from 3VDC to 5.5VDC. The MSP430G2231 will handle 1.8VDC to 3.6VDC.

The main power source is going to be the 12V accessory power circuit in a car. It seems logical to regulate the ~12VDC to 5VDC, and then use the regulated 5V to create the 3.3V needed for the MSP430. I was thinking about trying 3.3V on the backlight -- the specs state 3V typical, but don't specify a min or max value. I plan to see if I can just measure Vf of the backlight and add a current-limiting resistor to the circuit so I can safely use the 3.3V.

Anyhow, since I need two voltage regulators, does anyone have any recommendations for an ideal configuration? It looks like the LCD uses a max of 1.5mA @ 5V, and the backlight takes a max of 45mA. The MAX232 clone uses a max of 1mA @ 3.3V with no load (I have no idea what it would be under load), and the MSP430 looks like it uses about 4.2mA @ 3.3V and 16MHz. I think these are worst-case conditions.

Would you use a switching regulator to go from 12V to 5V, and then a linear regulator to go from 5V to 3.3V? Or one of the really low-count options as in the 3V Tips 'n Tricks from Microchip like a single zener or multiple diodes? I picked up some MC34063s to play with, but am open to anything else with a lower / smaller part count.

LCD specs I used:

enter image description here

MSP430 specs:

enter image description here

MAX232 specs:

enter image description here

I would sure appreciate any comments, like recommended approaches, or if I have flaws in my interpretation of the specs.

  • 2
    \$\begingroup\$ You only state a car battery, but car power has a number of very tricky power situations and surges that you need to design for. Read a number of articles on the web, since you'll need to include a lot of safety in the circuit and protection diodes. \$\endgroup\$ – kenny Apr 12 '11 at 10:17
  • \$\begingroup\$ @kenny is that true even for accessory power? I thought that circuit was relatively safe. I'll revise my post because I should have stated this, but I will also look into it more on the web. Thanks for the tip! \$\endgroup\$ – Dave Apr 12 '11 at 13:22

If power requirements are low (tens of millamps) I tend to use linear regulators. It doesn't matter whether you go from 12V to 3.3V immediately, or via 5V, the efficiency is the same (and rather bad). Make sure, however, if you use a regulator to go from 5V to 3.3V to use an LDO; the voltage difference is too small for standard regulators.

If you want to use SMPSs you'll have to look at the efficiency. This is in general lower for lower output voltages, and also if the Vin/Vout ratio gets higher. Let's say you go from 12V to 5V at 90%, from 12V to 3.3V at 80% and from 5V to 3.3V at 95%. So going from 12V to 3.3V in 1 step will be 80% efficient, if you do this in 2 steps (via 5V) your overall efficiency will be 90% * 95% = 85.5%. You'll have to make the calculation for your specific regulator.


Use +5V for the backlight. Select a resistor to drop 2V at 30 mA. (67 ohms). The backlight will get a nicely regulated current.

For the 3.3V, use a three-pin LDO. They are not expensive.

Be safe when connecting the LCD's output to the MSP430 input. Use a resistor divider or such. TI appnote.

  • \$\begingroup\$ 67 ohms, wtf?? Why not 68, that's an E12 value? BTW, I prefer diodes over resistors to create this voltage drop. In case of resistors, the voltage drop will vary with the load. \$\endgroup\$ – stevenvh Apr 12 '11 at 9:57
  • \$\begingroup\$ @stevenvh Totally agree with 68 Ohms. But preferring diodes? A resistor will indeed cause the voltage drop to vary with the load - that's the point when trying to stabilise the current. \$\endgroup\$ – MikeJ-UK Apr 12 '11 at 12:03
  • \$\begingroup\$ @Mike - You're assuming nice, smooth load characteristics, like positive dynamic impedance everywhere. If you don't have these your "regulator" may show some strange behavior. You don't know what the load characteristics are, so why not choose for a solution which has a fixed voltage drop, instead of something variable? \$\endgroup\$ – stevenvh Apr 12 '11 at 12:34
  • \$\begingroup\$ @stevenvh - We're talking about an LED backlight here. It will have a low dynamic resistance above the knee so driving it with constant voltage will give unpredictable current. \$\endgroup\$ – MikeJ-UK Apr 12 '11 at 12:40
  • \$\begingroup\$ @Mike - Alright, I guess I was misled by OP's "I was thinking about trying 3.3V on the backlight -- the specs state 3V typical, but don't specify a min or max value.". I presumed he wanted a 3V regulated voltage. \$\endgroup\$ – stevenvh Apr 12 '11 at 13:43

One red LED will drop you from 5 to 3.3 for applications < 20mA. Its super-cheap AND tells you if the power is on.

  • 1
    \$\begingroup\$ A red LED will drop 0 volts at 0 current. And a sleeping MSP430 is pretty close to zero current. \$\endgroup\$ – markrages Apr 12 '11 at 6:41
  • \$\begingroup\$ And the MAX232? \$\endgroup\$ – Andrew Stone Apr 12 '11 at 6:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.